干旱区生态环境知识资源中心

Spectral reflectance for soils and vegetation of the Yarmouk basin were correlated with surficial soil properties and vegetation biomass and cover. The overall aim of the study was to identify bands suitable for assessing soil and vegetation as indices for land degradation and desertification. Results showed that vegetation was well separated from soils in the shortwave infrared wavelength at 1480 nm. For most sites, the differences in the bandwidths (in the range of 8.5 nm to 90 nm) did not improve the differentiation of vegetation types. For all wavelengths, stronger correlation values (maximum R-2 = 0.85) were obtained for vegetation cover when compared with biomass (maximum R-2 = 0.54). Soil spectral reflectance tended to increase with salinity, with maximum correlations obtained in the blue wavelengths (470 +/- 10 nm, 485 +/- 90 nm), followed by green and the NIR bands, where R-2 values were around 0.60. Comparing results from radiometer measurements with results obtained from ASTER image bands showed that correlations tended to decrease with decreased spatial resolution for the investigated soil properties. For all wavelengths, spectral reflectance of degraded soils was higher than that for natural vegetation and irrigated crops with partial surface cover. Results of the study showed that the use of remote sensing indices related to vegetation cover and soil salinity would be recommended to map the extent of land degradation in the study area and similar environments. However, spectral unmixing should be applied to improve the correlations between satellite remote sensing data and surficial soil properties.